We present a simple scheme for width-independent multiplicative approximation algorithms for generalized covering and packing problems. We 
then present our main contribution- a novel sampling and acceleration technique for this scheme, which enables near-linear time algorithms. 
Applications of this result include near-linear time and width-free multiplicative approximation algorithms for fractional packing and 
covering problems and for normalized covering semi-definite programming.